This invention relates to a preprocessor for moving picture encoding, and more particularly to a preprocessor to be inserted between a source of input pictures and a moving picture encoder designed in accordance with ISO/IEC Recommendation CD11172-2.
Usually, moving picture data are encoded for the purpose of data-compression. There are many conventional methods of data-compression for a motion color picture. For example, HVS(human visual system), MC(motion compensation), VLC(variable length coding), and DCT(frequency transformation by discrete cosine transform) are well used and well known methods. Of these well known methods, HVS and MC are briefly described. The color representation for each picture element consists of three components: Y(luminance) and two chrominance components Cb and Cr. The human visual system is most sensitive to the resolution of an image's luminance component, so the Y values are encoded in full resolution, and the two chrominance vales are, for example, averaged for each 2.times.2 neighbourhood picture elements for datacompression.
In MC, prediction from a preceding picture frame(or frames) is used for compression of a current picture frame. A sequence of moving pictures is composed of a series of picture frames, and in a high probability, data of a picture frame is similar to that of a preceding picture frame. Thus, data of a picture frame can be compressed by using data of a preceding picture frame.
In 1993, a Recommendation of an International Standard for the coded representation of moving pictures was issued from ISO/IEC as ISO/IEC 11172. Hereafter, this Recommendation will be called MPEG(Moving Pictures Expert Group) after the name of the group which prepared the Recommendation.
In MPEG, there are three kinds of pictures, intra-coded pictures(I-Pictures), predictive coded pictures(P-pictures), and bi-directionally predicted pictures(B-pictures). Data of a B picture are determined by interpolation of data of a preceding picture and a succeeding picture.
Thus, use of B-pictures is very efficient to obtain a high compression ratio while preserving good picture quality. But, if B-pictures are used, some re-ordering of the picture sequence is necessary before encoding. Since B-pictures are coded using bi-directional motion compensated prediction, they can only be encoded after a subsequent reference picture(an I or P-picture) has been encoded.
The present invention provides one of the best apparatus for re-ordering of the picture sequence.